CPSC 601.38: Project Brainstorming Session

1
CPSC 601.38Project Brainstorming Session

• Carey Williamson
• Department of Computer Science
• University of Calgary

2
PROJECT OVERVIEW

• A typical course project might involve
• design/build/obtain appropriate testbed,
  environment, or platform for your project
• extend/customize as needed
• obtain relevant data/measurements needed
• design suitable experiment clear goal, identify
  factors, levels, performance metrics
• obtain and present (new/interesting) results

3
Some Data Sets and Traces

• Web server access logs (1996)
• Web proxy access logs (1998)
• MPEG video traces (20 x 40,000 frames)
• ISP measurements (4 traces, 1-2 minutes)
• FrameRelay/ATM traces (5 traces)
• Bellcore Ethernet LAN trace (1989)
• TCP/IP packet traces (LBL, 24 hours, 1.8M)
• See also the Internet Traffic Archive

4
Some Available Simulators

• ATM-TN simulator (ATM cell-level)
• Clustered Web server simulator (dws)
• Web proxy caching hierarchies (Muda)
• Distributed Web proxy simulator
• IP-TN simulator (U of C)
• IP-TNE (emulator) (U of C)
• LBLs ns-2 simulator (TCP packet level)

5
Some Useful Tools

• Synthetic Web proxy workload generation
• Web client traffic model (mosaic, 1995)
• LRD traffic analysis (R/S, V-T, AC, etc)
• GUI for traffic modeling/analysis (synTraff)
• Wavelet-based traffic model (Ram)
• Synthetic MPEG video trace generation
• SimKit programming language (UofC)

6
Issues and Ideas

• Improving/extending ProWGen
• temporal locality document mods scaling
• Web proxy caching hierarchies
• Hierarchical vs distributed caching
• Web response time modeling
• Improving network TCP flow model (dws)
• Wavelet-based traffic forecasting
• Wavelength assignment in WDM networks

7
1. ATM-TN System Overview
Input Data Set
ATM-T
ATM MF
ATM-N
TMF
SimKit
ESS
WarpKit
SMTW
Output Data Set
workstation
UNIX
Report Generation Scripts
Hardware SPARC, KSR, SGI
Report
8
CBR Poisson Ethernet JPEG/MPEG Web
TCP/IP/AAL5 ABR
Traffic Models
ATM MF
TMF
Switch and Network Models
SimKit
ESS
WarpKit
WaiKit
SMTW
UNIX Operating System
X
Sequential UNIX Workstations (SGI, SPARC, DEC,
HP) Parallel SGI Power Challenge, SPARC 1000
9
2. Distributed Web Server Model
1
File Server
2
3
Cache
Dispatcher (Front End)
Manager
N
Object Store
Web Clients
Server Nodes
10
Server Parameters

• Num server nodes
• Mem cache size
• Disk cache size
• Cache replacement policy for each (LRU, LFU,
  SIZE, DUAL)
• Comm. latency
• Cache consistency

• Dispatch policy (DNS, RR, Redirect, Load)
• Request distribution policy (requests, bw, conns,
  affinity, ...)
• Server bandwidth
• Per-request bandwidth
• BW scaling model

11
Performance Metrics

• Load balancing
• requests
• bytes
• bandwidth
• connections
• clients
• Relative improvement versus RR, Rand, etc

• Cache performance
• document hit rate
• byte hit rate
• Comm. overhead
• Avg response time
• Avg inflation factor
• Others...

12
3. Web Proxy Caching Model
Web Servers
Web Clients
13
Hierarchical Proxy Caching Simulation Model
Web Servers
Web Clients
14
Factors and Levels

• Cache size
• Cache Replacement Policy
• Recency-based LRU
• Frequency-based LFU-Aging
• Size-based GD-Size
• Workload Characteristics
• One-timers, Zipf slope, tail index, correlation,
  temporal locality model

15
ProWGen Conceptual View
ProWGen Software
Input Parameters
Synthetic Workload
1
Z
a
c
L
16
Key Workload Characteristics

• One-timers (60-70 useless!!!)
• Zipf-like document referencing popularity
• Heavy-tailed file size distribution (i.e., most
  files small, but most bytes are in big files)
• Correlations (if any) between document size and
  document popularity (debate!)
• Temporal locality (temporal correlation between
  recent past and near future references) Mahanti
  et al. 2000